Konstanz, Univ., Bachelorarb., 2004.

Every day, an overwhelming volume of new information is produced. Information Retrieval systems play a crucial role in managing this deluge and facilitate users to find relevant information. Simultaneously, the volume of scientific literature is also rapidly increasing, requiring powerful retrieval tools in this domain. Current methods of Information Retrieval employ language models based on the Transformer architecture, called Transformer-Encoder models, in this work. These models are generally trained in two phases: initially, a pre-training on general natural language data is performed, then fine-tuning follows, which adapts the model to a specific task such as classification or retrieval. Since Transformer-Encoder models are pre-trained on general natural language corpora, they perform well on these documents. However, scientific documents exhibit different features. The language in these documents is characterized by mathematical notation, such as formulae. Applying Transformer-Encoder models to these documents results in a low retrieval performance (effectiveness). A possible solution is to adapt the model to the new domain by further pre-training on a data set originating in the new domain. This process is called Domain-Adaptive Pre-Training and has been successfully applied to other domains. Mathematical Answer Retrieval involves finding relevant answers from a large corpus for mathematical questions. Both the question and the answers can contain mathematical notation and natural language. To retrieve relevant answers, the model must 'understand' the problem specified in the question and the solution of the answers. This property makes the task of Mathematical Answer Retrieval well suited to evaluate whether Transformer-Encoder models can model mathematical and natural language in conjunction. Transformer-Encoder models showed a low performance on this task compared to traditional retrieval approaches, which is surprising given the success of Transformer-Encoder models in other domains. This thesis, therefore, deals with the domain-adaption of Transformer-Encoder models for the domain of mathematical documents and the development of a retrieval approach using these models for Mathematical Answer Retrieval. We start by presenting a retrieval pipeline using the Cross-Encoder setup, a specific architecture of applying Transformer-Encoder models for retrieval. Then, we enhance the retrieval pipeline by adapting the pre-training schema of the Transformer-Encoder models to capture mathematical language better. Our evaluation demonstrates the strengths of the Cross-Encoder setup using our domain-adapted Transformer-Encoder models. In addition to these contributions, we also present an analysis framework to evaluate what knowledge of mathematics the models have learned. This analysis framework is used to study Transformer-Encoder models before and after fine-tuning for mathematical retrieval. We show that Transformer-Encoder models learn structural features of mathematical formulae during pre-training but rely more on other superficial information for Mathematical Answer Retrieval. These analyses also enable us to improve our fine-tuning setup further. In conclusion, our findings suggest that Transformer-Encoder models can be applied as a suitable and powerful approach for Mathematical Answer Retrieval.

Radiology departments utilize various visualization techniques of patients&rsquo / bodies, and narrative free text reports describing the findings in these visualizations are written by medical doctors. The information within these narrative reports is required to be extracted for medical information systems. Turkish is an highly agglutinative language and this poses problems in information retrieval and extraction from Turkish free texts. In this thesis one rule-based and one data-driven alternate methods for information retrieval and structured information extraction from Turkish radiology reports are presented. Contrary to previous studies in medical NLP systems, both of these methods do not utilize any medical lexicon or ontology. Information extraction is performed on the level of extracting medically related phrases from the sentence. The aim is to measure baseline performance Turkish language can provide for medical information extraction and retrieval, in isolation of other factors.

Thesis (M.Comp.Sc.)--University of Wollongong, 2006. / Typescript. Includes bibliographical references: leaf 140-157.

Thesis (Ph.D.) -- Mississippi State University. Department of Electrical and Computer Engineering. / Title from title screen. Includes bibliographical references.

Eidgenössische Techn. Hochsch., Diss.--Zürich, 2003.

Esta pesquisa discute a participação do bibliotecário na formação de equipes multidisciplinares dos grupos de Avaliação de Tecnologias em Saúde (ATS), caracterizando sua atuação num novo campo que se abre para os bibliotecários em instituições de pesquisa. O objetivo geral baseia-se na criação de uma Biblioteca Digital (BD) com os parâmetros de qualidade da informação inerentes a ATS, a partir dos documentos gerados pelo Serviço de Comutação Bibliográfica (SCB) da Rede de Bibliotecas da FIOCRUZ, recomendando a inserção do bibliotecário na equipe multidisciplinar para ATS. A metodologia foi dividida em três partes: levantamento do estado da arte do conhecimento produzido na Saúde Coletiva, onde se insere a Avaliação de Tecnologias em Saúde, e da Ciência da Informação, pesquisa exploratória com uma abordagem qualitativa para coleta de dados junto ao grupo de pesquisadores de ATS de diversas instituições públicas e privadas e uma abordagem quantitativa para coleta de dados dos profissionais do SCB da Rede de Bibliotecas da Fiocruz e análise dos dados. Verificou-se que existe uma participação ativa do bibliotecário nas atividades de ATS, no que diz respeito, à formulação de estratégias de busca em base de dados, revisão de protocolos de busca, localização de publicações relevantes, auxílio para realização de revisões sistemática para os grupos de pesquisa. Pressupõe a criação de uma BD permitindo o compartilhamento de todos os documentos digitais gerados pelas bibliotecas. Com essa iniciativa pretende-se contribuir para impulsionar a produção do conhecimento científico e tecnológico na área da saúde e de ATS. / This research discusses the librarians participation in the formation of multidisciplinary groups of Health Technology Assessment (HTA), characterizing its performance in a new field that opens to librarians in research institutions. The main objective is based on the creation of a Digital Library (DL) with quality parameters of information related to HTA, from documents created by the service of bibliographic Library Network Fiocruz, showing the necessity of the librarian in the multidisciplinary group to HTA. The methodology was divided into three parts: data collection of the art of knowledge produced in Collective Health, where the Evaluation of Health Technologies, and Information Science is, exploratory research with a qualitative approach to data collection with the group of HTA researchers from several private and public institutions and a qualitative approach to data collection of the Library Network Fiocruz professionals and data analysis. It was noted that there is an active participation of the librarians in HTA activities, formulating database search strategies, reviewing search protocols, locating relevant publications, helping locating systematic reviews to the search groups. This preuposes creation of a DL sharing all the digital documents created by the libraries. This initiative aims to contribute to boosting the production of scientific and technological knowledge in the health and HTA.

Made available in DSpace on 2014-12-17T14:56:05Z (GMT). No. of bitstreams: 1 MiriamVM.pdf: 955944 bytes, checksum: 0703348b1d68a5f32e8b6da9536de285 (MD5) Previous issue date: 2005-12-20 / Conselho Nacional de Desenvolvimento Cient?fico e Tecnol?gico / Attacks to devices connected to networks are one of the main problems related to the confidentiality of sensitive data and the correct functioning of computer systems. In spite of the availability of tools and procedures that harden or prevent the occurrence of security incidents, network devices are successfully attacked using strategies applied in previous events. The lack of knowledge about scenarios in which these attacks occurred effectively contributes to the success of new attacks. The development of a tool that makes this kind of information available is, therefore, of great relevance. This work presents a support system to the management of corporate security for the storage, retrieval and help in constructing attack scenarios and related information. If an incident occurs in a corporation, an expert must access the system to store the specific attack scenario. This scenario, made available through controlled access, must be analyzed so that effective decisions or actions can be taken for similar cases. Besides the strategy used by the attacker, attack scenarios also exacerbate vulnerabilities in devices. The access to this kind of information contributes to an increased security level of a corporation's network devices and a decreased response time to occurring incidents / Ataques a dispositivos conectados em rede constituem um dos principais problemas relacionados ? confidencialidade das informa??es sens?veis e ao correto funcionamento dos sistemas de computa??o. Apesar da disponibilidade de ferramentas e de procedimentos que dificultam ou evitam a ocorr?ncia de incidentes de seguran?a, dispositivos de rede s?o atacados com sucesso utilizando-se estrat?gias aplicadas em eventos anteriores. O desconhecimento dos cen?rios nos quais esses ataques ocorreram contribui de maneira efetiva para o sucesso de novos ataques. O desenvolvimento de uma ferramenta que disponibilize esse tipo de informa??o ?, ent?o, de grande relev?ncia. Este trabalho apresenta um sistema de apoio ? gest?o de seguran?a corporativa para o armazenamento, a recupera??o e o aux?lio na composi??o de cen?rios de ataque e informa??es relacionadas. Se um incidente de seguran?a ocorrer em uma corpora??o, o especialista em seguran?a deve acessar o sistema para armazenar o cen?rio de ataque espec?fico. Este cen?rio, disponibilizado atrav?s de acesso controlado, deve ser estudado para que decis?es efetivas possam ser tomadas em casos semelhantes. Cen?rios de ataque evidenciam, al?m da estrat?gia utilizada pelo atacante, vulnerabilidades existentes em dispositivos. O acesso a este tipo de informa??o contribui para a eleva??o do n?vel de seguran?a dos dispositivos de rede de uma corpora??o e para a diminui??o do tempo de resposta ao incidente ocorrido

Esta pesquisa discute a participação do bibliotecário na formação de equipes multidisciplinares dos grupos de Avaliação de Tecnologias em Saúde (ATS), caracterizando sua atuação num novo campo que se abre para os bibliotecários em instituições de pesquisa. O objetivo geral baseia-se na criação de uma Biblioteca Digital (BD) com os parâmetros de qualidade da informação inerentes a ATS, a partir dos documentos gerados pelo Serviço de Comutação Bibliográfica (SCB) da Rede de Bibliotecas da FIOCRUZ, recomendando a inserção do bibliotecário na equipe multidisciplinar para ATS. A metodologia foi dividida em três partes: levantamento do estado da arte do conhecimento produzido na Saúde Coletiva, onde se insere a Avaliação de Tecnologias em Saúde, e da Ciência da Informação, pesquisa exploratória com uma abordagem qualitativa para coleta de dados junto ao grupo de pesquisadores de ATS de diversas instituições públicas e privadas e uma abordagem quantitativa para coleta de dados dos profissionais do SCB da Rede de Bibliotecas da Fiocruz e análise dos dados. Verificou-se que existe uma participação ativa do bibliotecário nas atividades de ATS, no que diz respeito, à formulação de estratégias de busca em base de dados, revisão de protocolos de busca, localização de publicações relevantes, auxílio para realização de revisões sistemática para os grupos de pesquisa. Pressupõe a criação de uma BD permitindo o compartilhamento de todos os documentos digitais gerados pelas bibliotecas. Com essa iniciativa pretende-se contribuir para impulsionar a produção do conhecimento científico e tecnológico na área da saúde e de ATS. / This research discusses the librarians participation in the formation of multidisciplinary groups of Health Technology Assessment (HTA), characterizing its performance in a new field that opens to librarians in research institutions. The main objective is based on the creation of a Digital Library (DL) with quality parameters of information related to HTA, from documents created by the service of bibliographic Library Network Fiocruz, showing the necessity of the librarian in the multidisciplinary group to HTA. The methodology was divided into three parts: data collection of the art of knowledge produced in Collective Health, where the Evaluation of Health Technologies, and Information Science is, exploratory research with a qualitative approach to data collection with the group of HTA researchers from several private and public institutions and a qualitative approach to data collection of the Library Network Fiocruz professionals and data analysis. It was noted that there is an active participation of the librarians in HTA activities, formulating database search strategies, reviewing search protocols, locating relevant publications, helping locating systematic reviews to the search groups. This preuposes creation of a DL sharing all the digital documents created by the libraries. This initiative aims to contribute to boosting the production of scientific and technological knowledge in the health and HTA.

The integration of renewable energy sources (RES) into local energy distribution networks becomes increasingly important. Renewable energy highly depends on weather conditions, making it difficult to maintain stability in such networks. To still enable efficient planning and balancing, forecasts of energy supply are essential. However, typical distribution networks contain a variety of heterogeneous RES installations (e.g. wind, solar, water), each providing different characteristics and weather dependencies. Additionally, advanced meters, which allow the communication of final-granular production curves to the network operator, are not available at all RES sites. Despite these heterogeneities and missing measurements, reliable forecasts over the whole local distribution network have to be provided. This poses high challenges on choosing the right input parameters, statistical models and forecasting granularity (e.g. single RES installations vs. aggregated data). In this paper, we will discuss such problems in energy supply forecasting using a real-world scenario. Subsequently, we introduce our idea of a generalized optimization approach that determines the best forecasting strategy for a given scenario and sketch research challenges we are planning to investigate in future work.